Export
SWE-2
/wiki/swe-2
The link always opens the latest version of this page.
Ask an AI about it
Open in ChatGPTA new chat that reads this entry from the site Open in ClaudeThe same, in Claude
Sending this to someone?
MarkdownPlain text for notes apps and AI chats
WordA document to edit and annotate
PDFA4, for printing and sharing
ZIPEverything: the documents, the data, and instructions for AI assistants
Share the link instead of a file, and they will always see the latest version.
Files are a snapshot of today. The live page has anything newer.
stub ·
1 feed
SWE-2
Advanced coding model by Cognition achieving 50.0% on FrontierCode 1.1, 64% cheaper than Fable 5.1
Snapshot
- Vendor
- human cognition 1 source
- Category
- agentic coding model 1 source
Intelligence coverage
SWE-2 is an advanced coding model developed by Cognition that achieves 50.0% on FrontierCode 1.1 Main, 73.0% on DeepSWE 1.1, and 92.8% on Terminal-Bench 2.1 1.
The model is available across multiple platforms: Devin Desktop, Devin Web, Devin CLI, and Fusion 12. SWE-2 uses proprietary weights and does not support local deployment 2.
Facts 2
| Attribute | Value | Confidence | Sources | Cited |
|---|---|---|---|---|
| Vendor | human cognition | low | 1 | 2 |
| Category | agentic coding model | low | 1 | 2 |
Version 2
- 50.0% on FrontierCode 1.1 Main, 73.0% on DeepSWE 1.1, 92.8% on Terminal-Bench 2.1
- proprietary weights, no local run
Platform 2
Devin Desktop, Devin CLI, Devin Web, FusionDevin Desktop, Devin Web, Devin CLI, Fusion
Official links
No official links recorded yet.
Work here? Claim this page with a verified company email or DNS record.
Frequently mentioned together 4
Sources 2
- Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra · Hacker News (YC)
- Cognition's SWE-2 achieves 92.8 on Terminal-Bench 2.1 · Hacker News (YC)